These days, the group is again arguing about "data availability, ordering, finality," and as the number of terms increases, everyone starts pretending to understand... I’ll focus on one main thread: who do you trust, and when can you trust. Whether data can be "seen and retrieved" by others determines if you're accounting in the air; who queues up to determine ordering decides if you're being cut in line; finality is whether this matter will be overturned, so don’t celebrate prematurely before the verdict is final.

Recently, AI Agents/automated trading systems have also been quite popular, with narratives being hyped up, but I care more about the security boundaries of their on-chain interactions: giving too much permission is like handing over wallet keys to a "poetry-writing script." My current approach is very "patchwork": sign as little as possible, lock the limits first, review the contract carefully, and if necessary, run a small test first... Anyway, don’t be scared by the terminology; first, clarify the team and the decision time.